Pradeep Kumar Choudhury
Transcript of Pradeep Kumar Choudhury
Institute for Studies in Industrial Development4, Institutional Area Phase II, Vasant Kunj, New Delhi - 110 070
Phone: +91 11 2676 4600 / 2689 1111; Fax: +91 11 2612 2448E-mail: [email protected]; Website: http://isid.org.in ISID ISID
The Institute for Studies in Industrial Development (ISID), successor to the Corporate Studies Group (CSG), is a national-level policy research organization in the public domain and is affiliated to the Indian Council of Social Science Research (ICSSR). Developing on the initial strength of studying India’s industrial regulations, ISID has gained varied expertise in the analysis of the issues thrown up by the changing policy environment. The Institute’s research and academic activities are organized under the following broad thematic areas: Industrialization; Corporate Sector; Trade, Investment and Technology; Regulatory Mechanism; Employment; Public Health; Media Studies; and Other Issues.ISID has developed databases on various aspects of the Indian economy, particularly concerning industry and the corporate sector. It has created On-line Indexes of 203 Indian Social Science Journals (OLI) and 18 daily English Newspapers. More than one million scanned images of Press Clippings on diverse social science subjects are available online to scholars and researchers. These databases have been widely acclaimed as valuable sources of information for researchers studying India’s socio-economic development.
The Public Health Foundation of India (PHFI) is a public private initiative that has collaboratively evolved through consultations with multiple constituencies including Indian and international academia, state and central governments, multi & bi-lateral agencies and civil society groups. PHFI is a response to redress the limited institutional capacity in India for strengthening training, research and policy development in the area of Public Health.
Structured as an independent foundation, PHFI adopts a broad, integrative approach to public health, tailoring its endeavours to Indian conditions and bearing relevance to countries facing similar challenges and concerns. The PHFI focuses on broad dimensions of public health that encompass promotive, preventive and therapeutic services, many of which are frequently lost sight of in policy planning as well as in popular understanding.
About the PHFI
About the ISID
ISID-PHFI Collaborative Research Programme
PARENTAL EDUCATION AND INFANT MORTALITY IN INDIA:
Understanding the Regional Differences
November 2013
Working PaperNo: 157
Pradeep Kumar Choudhury
ISID-PHFI Collaborative Research Programme
ISID‐PHFI Collaborative Research Programme
PARENTAL EDUCATION AND INFANT MORTALITY IN INDIA: Understanding the Regional Differences
Pradeep Kumar Choudhury
Institute for Studies in Industrial Development
4, Institutional Area, Vasant Kunj Phase II, New Delhi ‐ 110 070
Phone: +91 11 2676 4600 / 2689 1111; Fax: +91 11 2612 2448
E‐mail: [email protected]; Website: http://isid.org.in
November 2013
ISID
Working Paper
157
© Institute for Studies in Industrial Development, 2013
ISID Working Papers are meant to disseminate the tentative results and findings obtained from the ongoing research activities at the Institute and to attract comments and suggestions which may kindly be addressed to the author(s).
CONTENTS
Abstract 1
1. Introduction 2
2. Past Research 3
3. Data and Methods 5
3.1 Data 5
3.2 Methods 7
4. Results and Discussion 12
4.1. Regional Variation of IMR by Parental Education 12
4.2. Role of Parental Education in the Regional variation of IMR:
Logit Estimates 16
5. Conclusions 25
References 27
Appendix 30
List of Tables
Table‐1 Infant mortality rates by parent’s education in EAG
and Non‐EAG states of India 13
Table‐2 Infant mortality rates by parents education and region
in EAG and Non EAG states of India 15
Table‐3 Logit Estimates of Infant Death by Parental Education
(Model 1 to 3) 17
Table‐4 Logit Estimates of Infant Death by Parental Education
(Model 4) 18
Appendix‐1 Notation and Definition of the Variables used
in the Logit Model 30
Appendix‐2 Summary Statistics 32
List of Figures
Figure‐1 Infant Mortality Rate by Parents Education and Region 16
PARENTAL EDUCATION
AND INFANT MORTALITY IN INDIA:
Understanding the Regional Differences
Pradeep Kumar Choudhury
[Abstract: Using data from the National Family Health Survey (2005‐06), this study examines the effect of
parental education in the regional variation of infant mortality in India. Although, research evidences show
that mother’s educations have a strong effect on reducing the mortality of young children, systematic attempts
to understand the role of parental education in the regional variation, are limited. Similarly, there is hardly
any attempt to examine the impact of mother’s exposure to mass media and her socioeconomic empowerment
(factors that are closely related to the education) on the risk of infant mortality in the regional level. Thus, the
need for this study lies with the argument that the role of parent’s education and other related factors in
reducing infant mortality differ significantly with the region, classified here as Empowered Action Group
(EAG)‐Non‐EAG states, and rural‐urban. While the overall infant mortality is 57 in major states of India, the
analysis shows that it varies enormously by parental education and regions. The regression results show that
both mother’s and father’s education are significantly associated in reducing the infant mortality across the
regions and major states of India, although the relative effect of different levels of education of the parents
varies between EAG‐Non‐EAG states and rural‐urban regions. Similarly, it is also evident that the children
born to the mothers having any kind of exposure to the mass media have lower probability of death in their
infant stage compared to the children born to the mothers having no mass media exposure and it works more
effectively in the regions that are underdeveloped such as EAG states and rural areas. The analyses of the
findings suggest that the low level of female education is a major hindrance to reduce the infant death in rural
India and EAG states than that of urban India and Non‐EAG states respectively. The results are very robust
to different potential confounding factors including socio‐economic, demographic, accessibility to health care,
and sanitation related variables. The policy implication of the study include, obviously, providing education to
the parents, particularly to the mothers of backward regions or states. Besides parent’s education, attempt
should also be made to increase the scope of getting mass media exposure and higher level of socioeconomic
empowerment of the mother to reduce the infant mortality in India.] Keywords: Parental Education, Infant Mortality; Regional Variation; India
Assistant Professor, Institute for Studies in Industrial Development, New Delhi. E‐mail:
Acknowledgement: I am thankful to Dr Jagannatha Behera from CSRD, JNU, Dr Indranil Mukhopadhyay
from PHFI, and faculty members from ISID for their valuable suggestions. However, the usual
disclaimers apply.
2
1. Introduction
Infant Mortality Rate (IMR) which is defined as the deaths of infants of age less than one year
per thousand live births is considered as a sensitive indicator of living and socio‐economic
conditions of a country. The high IMR of a country indicates the unmet health needs,
unfavourable environmental factors, and low health and socio‐economic status of its
population. It is also considered as one of the indicators under the Millennium Development
Goal (MDG) four i.e., to reduce the child mortality by two‐thirds between 1990 and 2015. The
latest mortality estimates for the year 2011 in India indicates an IMR of 44 which is higher
than the global figure of 40 in the corresponding year (Registrar General of India, 2011a;
World Health Statistics, 2012). Further, within India the infant mortality rate varies widely
across different regions and states. In 2011, the IMR is 48 in rural areas and 29 in urban areas.
Similarly, it varies from 12 in the state of Kerala to 59 in Madhya Pradesh. As per the latest
National Family Health Survey (NFHS‐3), the data base on which the present study is based
on, for all India level the IMR is reported to be 57 and it varies from 62 in rural areas to 41 in
urban areas and among the states, it is highest in Uttar Pradesh (73) and lowest in Kerala and
Goa (15).
What factors contribute to such variations is not being examined in detail in a cohesive
framework in India. However, the cross country literature, have found that the regional
diversity, socio‐economic and demographic factors can be the cause of such variations
(Wang, 2003). Quite a few studies have argued that education is the most influenced factor in
differentiating the infant and child mortality levels within all the socioeconomic factors
(Mondal et al, 2009). In the developing world, including India, there is now clear evidence of
differentials in child survival rates associated with the education of mothers (Ware, 1984;
Basu and Stephenson, 2005). Although majority of the scholars working in this domain have
analysed the pattern of infant and child mortality rate by mother’s education (Caldwell, 1979,
1994; Cladwell and McDonald, 1982; Desai and Alva, 1998; Gakidou et al, 2010; Papageorgiou
and Stoytcheva, 2012; Saikia et al, 2013), there is very little investigation on the impact of
father’s education and the aggregate level of education of the parents on it. Similarly, very
little is known about the relationship between mother’s mass media exposure and
socioeconomic empowerment on mortality and health status of the child. Though a couple of
studies have emphasized the rural‐urban disparities in the infant and child mortality rate and
the factors contributing to these in India (e.g., Saikia et al, 2013), perhaps no attempt is made
to examine the differences by EAG and Non‐EAG states, which is an important classification
made by the Government of India on the basis of crucial health indicators1.
In light of these caveats, this study examines the role of parental education, controlling for
other factors, on the infant mortality rate in India using NFHS‐3 data collected in the year
2005‐06. In the present phrase of development in India, examining these concerns are having
particular interest because 1) the movement towards universal health coverage instituted by 1 Detail is given in section 3 on data and methodology.
3
government of India targets for reducing health related mortality in the regional level by
strengthening the social determinants of health, including education 2) special effort is being
given in the health policy domain to bridge the regional gap in the infant mortality by
creating awareness through mass media and other informal education to the mothers, with a
special focus on rural region and backward states. For example, the programme on National
Rural Health Mission (NRHM) targets to reduce the IMR in rural region by involving local
communities in the process. The paper examines the regional aspect of this by grouping the
major states of India into EAG and non‐EAG states and also by dividing these states into
rural and urban region. The main purpose of the study is to ascertain the nature of cross‐
sectional (regional level) relationship of infant mortality rate with educational levels of the
parents. The importance of carrying out this study also lies with the widely argued fact that
the empirical studies on infant and child mortality at the regional level or geographic location
are more informative for designing health policies of a country.
The rest of the paper is organised as follows: The studies on the impact of parental education
and other related variables on infant mortality in India and rest of the world are reviewed in
section 2. Section 3 describes the data and methodology used in the study. Section 4 discuses
the empirical results. These include: (a) region wise differentials of IMR by parental
education; (b) examining the impact of mother’s and father’s education and other
socioeconomic and demographic factors on IMR in the regional level in India using logit
model. Section 5 concludes.
2. Past Research
How does education (particularly parental education) matter for infant and child mortality
and health status of the child? Quite a few number of studies (Calwell, 1979, 1994; Cladwell
and McDonald, 1982; Rama Rao et al, 1997; Desai and Alva, 1998; Das and Dey, 2003;
Khasakhala, 2003; Kravdal, 2004; Gakidou et al, 2010; Papageorgiou and Stoytcheva, 2012;
Saikia et al, 2013) in this domain have shown a negative relationship between maternal
education and infant and child mortality, although the amount of education required to
produce a significant reduction in mortality differs from culture to culture. However, given
the close link between education and other positive socio‐economic conditions, researchers
differ in their propensity to move beyond this correlation to argue that education causes low
infant and child mortality. It is well established in the literature that education is linked to
family socioeconomic condition, which in itself has an impact on child mortality and health
status. But above and beyond this, the parent’s education is expected to bring about certain
changes in individual behaviour that results a positive impact on child mortality and health.
Using the fixed effect model, the study by Desai and Alva (1998) shows a consistent negative
relationship between maternal education and the probability of infant death. However, the
result of the logistic regression, introducing direct controls for some of the socio‐economic
variables reduces the effect of maternal education on infant mortality. For example,
introducing controls for husband’s education and access to piped water and toilet attenuate
the impact of maternal education on infant mortality. Gakidou et al, (2010) have estimated the
4
contribution of improvements in women’s education to reduction in child mortality using the
data of 219 countries, gathered between 1953 and 2008. The coefficient for women’s education
implied that for every 1 year of increase in the education of women of reproductive age, the
child mortality decreased by 9.5 per cent. In south Asia, the expansion of women’s education
accounted for 39.1 per cent of the reduction in the number of child deaths.
Mother education emerges as the single most important determinant of child health‐care
utilisation in India when the influences of other intervening factors are controlled
(Govindasamy and Ramesh, 1997). The empirical results show that a higher level of maternal
education results in improved child survival because health services that effectively prevent
fatal childhood diseases are used to a greater extent by mothers with higher education than
by those with little or no education. Educated women with at least middle schooling are
nearly eight times more likely to receive antenatal care for their births than illiterate women,
and literate women with less than middle schooling are more than three times as likely.
Similarly, births to women who have completed middle school are five to eight times more
likely to receive maternal care as compared to births among illiterate women. Papageorgiou
and Stoytcheva (2012) have examined the relationship between female human capital
inequality and infant mortality using female‐specific education Gini coefficients for 108
countries, in addition to their average female education. The results show that higher
education inequality among women leads to substantially higher infant mortality. The study
suggests that if infant mortality reduction is a priority for policy makers, then educating the
least educated women is a priority. Kapoor (2010) has examined the socio‐economic
determinants of the reducing infant mortality rate in India using a panel dataset of 666
districts from the 1991 and 2001 census. The results of the Quantile regression technique
show that female literacy has a negative and statistically significant effect on infant mortality.
One per cent increase in female literacy is associated with a 23 per cent drop in IMR. Thus,
educating more females leads to significant reductions in IMR across the districts, though the
relative effect varies. The findings of this study clearly demonstrate the role of woman’s
empowerment (measured in terms of their educational level) in reducing infant mortality and
the results are strongest in districts with high levels of IMR.
Kiros and Hogan (2001) have examined the role of parental education in reducing excess
child mortality in Africa. Trends in child mortality by parental education for the period 1979‐
1994 shows that it is highest among children born to illiterate parents and decreases with
both mother’s and father’s education. The results of Poisson regression model show a
significant reduction in number of children dead with the parents having any level of
education compared to when both parents are illiterate. Even after controlling the additional
covariates like urban‐rural, food crisis and war intensity, the effect of parental education in
reducing number of child deaths remained consistent and strongly significant. Both mother’s
and father’s education are found to have independent significant impacts in reducing child
mortality. Child mortality is highest among children born to illiterate mothers and illiterate
fathers. Singh‐Manoux et al, (2008) have examined the association between adult education
and child mortality using NFHS‐2 data collected in 1998‐99. The results show that compared
5
to those with no education, 9 or more years of education for the head of household and the
spouse was associated with lower child mortality. The study has suggested that adult
education has a protective association with child mortality in India and the confounding
factors like caste, household wealth and urbanisation do not modify or completely attenuate
this association. Similarly, using NFHS‐1 data collected 1992‐93, Gunasekaran (2005) has
critically analysed the level, trends, differentials and determinants of infant and child
mortality in rural India. The multivariate analysis reveals that education of the mother has a
significant effect on infant mortality only in the high and low mortality group states. The
relative risk of infant mortality among the children of literate mothers was significantly less in
the high and low mortality groups as compared to illiterate mothers.
The literature reviewed here confirms that the parent’s education (specifically mother’s
education) helps to reduce infant and child mortality and this is true even after controlling
the confounding factors that are linked to education like occupation, economic status etc.
Though largely the impact of parent’s education on infant and child mortality is clear from
the literature, less is talked about its impact by region. Similarly, the handful of studies
available on the determinants infant and child mortality in India are narrowly focused by
examining the impact of mother’s education, without considering many other important
variables like father’s education, mother’s exposure to mass media, and socio‐economic
empowerment of the mother. Thus, this study may be an extension in this direction in a
systematic way.
3. Data and Methods
3.1 Data
The data for this study comes from the 2005‐06 National Family Health Survey (NFHS‐3), the
third and also the latest in the series of these national surveys. It was preceded by NFHS‐1 in
1992‐93 and NFHS‐2 in 1998‐99. The NFHS is a large scale, cross‐sectional, multi‐round
survey conducted in a nationally representative sample of households throughout India. The
NFHS surveys were conducted under the stewardship of the Ministry of Health and Family
Welfare, Government of India with the objective to provide national and state level estimates
of fertility, family planning, infant and child mortality, reproductive and child health,
nutrition of women and children, the quality of health and welfare services, and
socioeconomic conditions.
NFHS 3, the data set on which this paper is based on, interviewed a sample of 1,24,385
women age 15‐49 and 74,369 men age 15‐54 in 1,09,041 households from all 29 states. The
survey has used standardised questionnaires (household questionnaire, women’s
questionnaire, and men’s questionnaire), sample designs and field procedures to collect the
information and a detailed description of the survey design of NFHS‐3 is given in the
national report (IIPS and ORC Macro, 2007). The fieldwork was conducted in two phases
between November 2005 and August 2006. The individual women response rate, i.e., the
6
number of completed interviews per 100 eligible women identified in the households, was 95
per cent for all India (93% in urban areas and 96% in rural areas). Similarly, the response rate
for eligible men was 87 per cent for the country as a whole (85% in urban areas and 90% in
rural areas).
For the present analysis, the data is gathered from the household questionnaire and the
women questionnaire. Since the focus of this study is on infant mortality, the information on
the number of children who were born, survived, or died is of particular concern. The survey
had asked all women age 15‐49 to provide a complete history of their births including for
each live birth, the sex, month and year of birth, survival status, and age at the time of the
survey or age at death. Age at death was recorded in days for children dying in the first
month of life, in months for other children dying before their second birthday, and in years
for children dying at later ages. The data on birth history allows for estimates of IMR for all
India and also in the regional level. The women questionnaire was also designed to provide
their background characteristics, antenatal, delivery and post natal care, and child
immunization and feeding practices, some of which are used as exposure variables in the
analysis. Information on some of the socio‐economic variables like the availability of toilet
facilities, source of drinking water etc. used in the analysis were drawn from the household
questionnaire.
The base sample for the present study constitutes 57,175 children born in the last five years
preceding the survey (i.e. in the period of 2000‐2005) in major states of India2. These were
children born to ever‐married women aged 15‐49 for whom the information was collected in
the survey. Out of the total children born during this period, 3,269 children have died before
their first birthday. Thus, the infant mortality rate turned out to be 57 per 1,000 live births,
which is considered as the dependent variable for the empirical analysis. It is a dummy
variable and takes value 1, if the infant is died and 0, if the infant is alive.
Besides analysing the differences of IMR exist in rural and urban regions by parental
education, the study has also examined the disparities between EAG and Non‐EAG states of
India. The EAG states were formed in the Ministry of Health and Family Welfare (MoHFW)
in March 2001 especially to ensure population stabilisation and intersectoral convergence and
these are the states with high fertility rates and weak socio‐demographic indicators. The
2 The major states (22 in total) include Andhra Pradesh, Assam, Bihar, Chhattisgarh, Delhi, Goa,
Gujarat, Haryana, Himachal Pradesh, Jammu and Kashmir, Jharkhand, Karnataka, Kerala, Madhya
Pradesh, Maharashtra, Odisha, Punjab, Rajasthan, Tamil Nadu, Uttarakhand and Uttar Pradesh, and
West Bengal. Seven North‐Eastern states excluded from the analysis are Arunachal Pradesh,
Manipur, Meghalaya, Mizoram, Nagaland, Sikkim and Tripura. These states are having some
geographical disadvantages as they are located in hilly region and also concentrated with tribal
population. Thus, the government of India treats this region separately when making the policies and
programmes of the country. These states consist only about 1.2 per cent of India’s population as per
2011 census.
7
objective of classifying EAG states were to facilitate the creation of area‐specific programmes,
with special emphasis on eight states that have been lagging behind in containing population
growth to manageable limits. These include eight states namely Bihar, Chhattisgarh,
Jharkhand, Odisha, Madhya Pradesh, Rajasthan, Uttarakhand, and Uttar Pradesh. However,
Assam (another north‐eastern state) is included to it as it is having similar characteristics. The
analysis in this case examines the concern of ‘Does the impact of parental education on IMR
differ in the sates with different socio‐economic settings?’ The sample for the EAG and Non‐
EAG states are 33,073 and 24,102 children respectively born in the period 1‐59 months
preceding the survey. Similarly, the sample sizes in the rural and urban regions are 42,533
and 14,642 children respectively.
3.2 Methods
The analysis is carried out in two levels. First, the levels of infant mortality rate is presented
by different levels of education of the mother, father, and parents (educational attainment of
father and mother taken together) for major states of India, and also two groups of states
(EAG and Non‐EAG). In each of these regions the pattern of infant mortality rate is discussed
by place of residence (rural‐urban). It will provide a descriptive picture on the issue. Second,
four different binary logistic regression models are fitted to examine the impact of parental
education on infant death. The first three models examine the relationship between infant
death and parental education without controlling for other variables. The first model
examines the effect of mother’s education alone, model 2 analyses the impact of father’s
education alone whereas the third model looks at the effect of both father and mother
education on the risk of infant death.
Model 4 extends model 3 through the addition of potential confounding factors including
mother’s exposure to mass media, socio‐economic empowerment of the mother, economic
status of the household, working status of the mother, drinking water and sanitation facility,
accessibility to health care, mother’s age, birth order of the child, size of the household, and
caste. Each of these models is estimated for EAG and Non‐EAG states, and rural‐urban
locations, to examine the role of parental education on infant mortality by region. In the
literature, it has been argued that education is one of many indices of socio‐economic status
and that the strong positive relationship between education and infant and child mortality
rate is merely a reflection of the fact that educated parents come from wealthier families, live
in urban settings where health care is more accessible. Thus, controlling for the possible
impact of other socio‐economic variables is an important part of the exercise to determine if
the positive impact of education on infant and child mortality is robust. Notation and
definition, and the summary statistics of explanatory variables used in the analysis are given
in Table A1 and Table A2 respectively, in appendix.
Inclusion of different explanatory variables in the logit model is elucidated here in detail.
8
Parent’s Education: One of the primary interests of the study is to empirically estimate the
effect of parent’s education on infant mortality rate. It is observed that the parents with lower
level of education are having higher infant mortality rate. Quite a few number of studies have
examined the impact of mother’s education on IMR assuming that mother’s education is
more important in reducing IMR comparison to father’s education. Mothers having higher
educational attainment can care their new born child well than the mothers having lower
level education, which declines the IMR. Similarly, educated mother are aware of the proper
antenatal care i.e. the regular medical and nursing care recommended for women
during pregnancy, which has significant impact on reducing IMR. It is expected that the
parents with higher level of education prefer institutional delivery, do not follow taboos and
superstitions in caring their baby, which helps in reducing the IMR. More clearly, parents
with higher level of education are generally better informed about the antenatal and
postnatal care than the parents with less level of education. To examine this fact, education
level of both father and mother are used in the analyses separately, which are classified into
four groups: (a) illiterate or having no education; (b) elementary level of education i.e., having
1 to 8 years of schooling3; (c) secondary level of education i.e., having 9 to 10 years of
schooling; and (d) senior secondary and above level of education ‐ having 11 or more years of
schooling.
Mother’s Exposure to Mass Media: Higher exposure to the mass media of the mother may be
regarded as an important factor in reducing the IMR. Mothers who are listening radio,
reading newspapers, watching television, participating in street plays etc. are more aware
about their children’s health care compared to the mothers without this knowledge. The
informal learning of the mothers from mass media (mainly through advertisements and
public awareness campaigns) helps for better antenatal and postnatal care. It may be
expected that the mass media is an important source of information on health related issues
even for the illiterate mothers. The mother’s exposure to mass media in the present analysis
include their involvement in three different mass media instruments like reading
newspapers, watching television and listening radio. The mass media exposure is taken as a
dummy variable, 1 if the mother listen radio, read newspaper and watch television at least
once a week, and 0, if mother does not listen radio, does not read newspaper and does not
watch television even once in a week.
Socio‐Economic Empowerment of the Mother: It is expected that the socio‐economic
empowerment of the mother may have a significant effect in determining the IMR. This is
mainly because the mother having higher socio‐economic empowerment have greater
chances of taking decision about their health care and also on their children’s health care. For
example, they can decide to take proper antenatal care and also to deliver their children in the
3 Taking 1 to 8 years of schooling (elementary level of education) as a category has a special
significance in the Indian context. This is considered as the compulsory level of education that must
be given free of cost to all citizens of the country.
9
hospital. Women who do not avail proper health care have a fair chance of delivering
premature baby, which leads to infant death. Hence, the mother’s socio‐economic
empowerment is included as one of the explanatory variables in the logit analysis. This is
calculated by considering the decision taken by the mother alone on four different socio‐
economic aspects. These include: decision on mother’s health care, decision on large and
daily household purchases, decision on spending husband’s money, and decision on visit to
family and relatives. The level of socio‐economic empowerment is classified as a dummy
variable, takes 1 if the mother has taken decision alone at least in one aspect, and 0, if mother
alone has not taken decision in any of the four aspects.
Household Economic Status: Economic status of the household is expected to matter
significantly in determining the level of IMR. Poor households may not afford the health care
(both ante antenatal and postnatal) properly, which leads to infant death. Contrary,
households with higher economic status able to avail better health care by spending a good
proportion of their income. Usually, rich households keep the expenditure on health in the
priority list of their household budget. Further, poor families go for more number of children
as they consider them as the future asset of their households, which is not the case among
rich families. To examine this, economic status of the household measured in terms of wealth
index is included as an explanatory variable in the logit model. The wealth index was
constructed from household‐level data, using Principal Components Analysis (PCA). The
input information for this analysis came from household ownership of items ranging from
furniture and vehicles; to dwelling characteristics such as water source, sanitation facilities,
and the home’s construction materials; and to whether a household member had a bank or
post office account4. The wealth quintiles were classified into lowest, second, middle, fourth
and highest in the original data file. However, for the present analysis it is categorised as poor
(considering lowest and second together), middle, and rich (taking fourth and highest
together).
Working Status of the Mother: Whether the mother is working or not may influence the level
of IMR. It is commonly perceived that the infant whose mother is working in the job market
has a higher chance of death, mainly because the mother does not get time to take care her
baby properly. Also, one apparent cost associated with formal work force participation by the
mothers of the infants is the necessary abandonment of breastfeeding, which is highly
recommended during the first year of the baby to reduce the chance of illness and
4 The wealth index is based on 33 assets and housing characteristics: household electrification; type of
windows; drinking water source; type of toilet facility; type of flooring; material of exterior walls;
type of roofing; cooking fuel; house ownership; number of household members per sleeping room;
ownership of a bank or post‐office account; and ownership of a mattress, a pressure cooker, a chair, a
cot/bed, a table, an electric fan, a radio/transistor, a black and white television, a colour television, a
sewing machine, a mobile telephone, any other telephone, a computer, a refrigerator, a watch or
clock, a bicycle, a motorcycle or scooter, an animal‐drawn cart, a car, a water pump, a thresher, and a
tractor.
10
consequently death. Contrary to this, the other possibility is that the probability of infant
death is less among working mothers. It is because they are usually educated and aware of
their child’s health. Further, due to their better economic status (compared to the mothers
who are not working) they can afford the health care facilities without much difficulty.
Working mothers can also hire some other women to take care of their babies when they go
for work. Thus, working status of the mother may have different effects upon the IMR. In the
present study, this is used as a dummy variable, i.e. 1, if the mother is working in the job
market, and 0, if the mother is not working, i.e. she is a housewife.
Caste5: Social category/caste is known to affect many aspects of life in India and is likely to
affect levels of infant mortality as well. In India, the social category is broadly classified into
four categories namely scheduled caste, scheduled tribe, other backward class, and
others/forward caste. Scheduled castes and scheduled tribes, and in some cases other
backward classes are castes that the government of India considers as socially and
economically backward and in need of special protection from injustice and exploitation. It is
generally expected that the children of scheduled caste and, in particular, children of
scheduled tribes are at a higher mortality risk compared to the children of other backward
classes and forward castes. Some of the effect of caste membership on mortality may be due
to the differences in life‐style based on traditions and beliefs. Such differences may include
customary practices related to childbirth, infant feeding, and health care, and these might
have an effect on infant mortality independent of other variables. To find out the effect of
caste on IMR, the dummy variable is used in the logit model and takes the value 1, if the
children belongs to forward caste or other backward classes families, and 0, otherwise i.e., if
the children belongs to scheduled caste or scheduled tribe families.
Sanitation: The sanitation condition of the household measured in terms of the availability of
toilet facilities may be regarded as an important determinant of IMR in India. It is expected
that the availability of toilet facility in the household can reduce the probability of infant
death significantly. To examine this fact, whether the household has a toilet facility or not is
included as a dummy variable in the logit model. It is important to note here that besides the
availability of toilet facilities, the type and condition of the same is also an important aspect in
reducing the IMR. But it is not being considered in the present analysis.
Drinking Water Facility: The quality of drinking water used by the mother and sometimes
by the infant plays a crucial role in determining the probability of infant death. Using safe
drinking water may reduce the infant death significantly. In the NFHS‐3 data respondents to
the household questionnaire were asked about the different source of drinking water for the
household. These include piped water, tube well water, drug well water and surface water.
However, in the present study, these are re‐classified into two categories namely piped water
5 Information on caste was obtained for the head of the household, and women were assumed to have
the same caste as the head of the household.
11
and others (tube well water, drug well water, and surface water taken together) and taken as
a dummy variable in the analysis.
Place of Delivery: Place of delivery is also an important determinant of child survival. In
many developing countries, including India, several children die owing to the lack of safe
delivery facilities. Specifically women from rural region and lower income groups deliver
their babies at home with the presence of relatives and neighbours without going for the
medical, which is risky both for the mother and the new born baby. On the other hand, it is
expected that the women going for institutional deliveries have access to trained medical
professional (doctors and nurses) who can take care of both mother and babies if there will be
any difficulty. Considering this, the present study has included the accessibility to
institutional delivery as a controlling factor in the analysis. It is a dummy variable and takes
the value 1, if the mother has gone for an institutional delivery and 0, otherwise.
Mother Age: Age of the mother is an important demographic factor which influences the
IMR. Typically, mortality risks are greater for children born to mothers who are too young or
too old. The probability of infant death is expected to be higher if the mother’s age is usually
less than 20 years and more than 30‐35 age. More clearly, other things being equal, a mother
with higher age (usually maximum of 30‐35 years old) giving birth to a baby will have lower
risk of dying compared to a mother with less than 20 years old and more than 35 years old.
This is due to the fact that, pregnancy in the teenage (normally less than 20 age) and over the
age of thirty five of the mother leads to preterm birth of the child which is an important cause
of infant death. Mother’s age is classified here into four categories namely, less than 20 years
old, 20‐29 years old, 30‐39 years old and 40‐49 years old. An attempt was made to include
‘mother’s age at the time of birth of her child’ instead of (‘mother’s age’) in the analysis, as the former
one is expected to be more important than latter one in determining IMR. However, it is found
to be highly correlated with ‘birth order of the child’ (included as another explanatory
variable in the analysis), hence dropped.
Birth Order of the Child: Birth order may also play a key role in the probability of infant
mortality, though the direction of the effect is a priori ambiguous. A number of studies
indeed point to a U‐shaped effect of birth order, with the probability of infant mortality
declining after the first child and increasing again for children of birth order four and higher
(Titaley et al, 2008 and Uddin and Hossain, 2008). First born children are likely to be raised by
parents with limited skills and experience, possibly increasing the risk of infant mortality.
Similarly, births of very high order may have mothers who are physically depleted at the
time of conception and throughout pregnancy. They are thus more likely than other children
to suffer from conditions associated with high mortality risk such as foetal growth retardation
and low birth weight. High‐order births are also born into families that already have a
number of young children who compete for resources and parental care (Pandey et al, 1998).
Contrary to this, according to the hypothesis of intra‐household resource competition, first
born children are more likely to capture vital resources such as food and care, thereby
reducing their mortality risk (Vos et al, 2004). On the other hand, it has been found that first
12
born children, who are more likely to be born to mothers at younger reproduction ages,
experience a higher mortality risk than children of a higher birth order. To account for this
effect, the present study has included the birth order of the child in four dummies
representing birth orders 2, 3, 4, and 5 and above ( 5) with birth order 1 as the reference
category.
Household Size: The size of the household is expected to have an impact on the infant
survival. Two contrary arguments on this can be posed here. First, the probability of infant
death may be higher in the large households. It is mainly because of the lower socio‐
economic status of these households, as large portion of the income will go for basic
consumption. They might not give due emphasis on health care and other related aspects.
Second, one can expect a lower probability of infant death in large households, as there will
be enough people to take care of the newly born child. To examine this, the total size of the
household (a continuous variable) is considered as an explanatory variable in the analysis.
4. Results and Discussion
4.1 Regional Variation of IMR by Parental Education
Table 1 confirms that higher the level of educational attainment of the mother, lower is the
level of infant mortality in India during 2000‐2005. The IMR among children born to illiterate
mothers was 68.5, which is about 2.5 times higher than those born to mothers with 11 or more
years of education. The IMR is 54.7 among the children born to mothers having elementary
level of education and 41.1 to mothers with secondary level of education. The decline in the
IMR is highest between secondary and senior secondary and higher level of education of the
mother, reduction of 14.9 points (41.1 to 26.2). It is important to note here that the IMR for the
children born to literate mothers is less than the national average which is 57 whereas it is
higher (68.5) among illiterate mothers. A mother’s education is important in reducing IMR
because it facilitates her integration into a society impacted by traditional customs, exposes
her information about better nutrition, use of contraceptives to space births and knowledge
about childhood illness and treatment (NIMS et al, 2012). It is also argued that education
makes a mother socially advanced, free from traditional values and changes her pattern of
behaving and attitude (Mondal et al, 2009). Education may also increase the autonomy of the
mother that is necessary to advocate for her child in the household and the outside world.
In India and other such developing societies, father’s education plays an important role in
earning income, which in turn ensures access to child health facilities. Thus one can expect a
direct relationship between father’s education and infant death. It is apparent from Table 1
that higher the educational attainment of the father lower is the level of infant mortality.
More clearly, children born to the father’s with senior secondary and above level of education
experience lowest IMR than that of father’s who have elementary and secondary educational
level. The large decline in the IMR is seen between the fathers having 1 to 8 years of
education and 9 to 10 years of education whereas the difference in the IMR between illiterate
13
fathers and father’s having elementary level of education is minimum. This may reflects the
dynamism involved in the education and labour market relationship in India. The scope of
getting employment and also the earning does not vary much between illiterate and persons
having primary level of education. However, there is not enough literature in supporting the
relationship between father education and IMR in India or otherwise, as the studies in this
domain are largely focused on mother’s education.
Table 1
Infant mortality rates by parent’s education in EAG and Non‐EAG states of India
Variables EAG States Non‐EAG States India
Mother Education
No education 74.0 54.4 68.5
1‐8 years complete 65.9 46.4 54.7
9‐10 years complete 51.2 33.9 41.1
> 10 years complete 32.6 22.8 26.2
Father Education
No education 74.1 52.2 66.6
1‐8 years complete 72.1 48.6 60.2
9‐10 years complete 60.0 38.4 49.1
> 10 years complete 46.5 28.7 37.5
Total 67.5 42.9 57.0
Source: Authors’ calculation from NFHS‐ 3 Data
While the overall IMR is 57 in major states of India, it varies widely between EAG and Non‐
EAG states (Table 1). As expected, the IMR is higher in EAG states (67.5) compared to Non‐
EAG states (42.9). This confirms the widely accepted argument that infant mortality is an
indicator of the socioeconomic wellbeing of a society. The high infant mortality in EAG states
signifies that there exists an unequal distribution of resources between these two regions and
thus, a special attention is needed to minimise this gap. Further, studies on infant mortality in
India and elsewhere have documented that the regions with lower socioeconomic settings
(particularly lack of parental education) are associated with high degree of son preference
which leads to excess infant mortality among female children due to preferential treatment in
feeding and medical treatment for sons (Arnold, Choe, Roy 1998; Williamson 1976).
Although, in both the regions mother’s education is inversely related with IMR, the rate of
decline by levels of education varies. With the increase in the mother’s level of education, the
IMR has declined fast in EAG states compared to non‐EAG states. For example, the decline in
the IMR is 18.6 and 16.7 points with the increase in the years of education from 1‐8 years to 9‐
10 years and then more than 10 years of education respectively in EAG states. The respective
figures are 12.5 and 11.1 in Non‐EAG states. Similarly, educational attainment of the father
has an inverse relationship with the IMR in both EAG and Non‐EAG states. The pattern in
the decline of IMR with father’s education is more or less same with the pattern of mother’s
education, except the case of no education to elementary level of education. The pattern of
14
infant death by parental education in EAG and Non‐EAG states (Table 1) spells out the
following: first, the role parental education in reducing the infant death differs with the
socioeconomic settings of the regions and interestingly, the reduction is higher in the
underdeveloped region than that of developed region. Second, irrespective of region, getting
compulsory level of education by the mother helps to reduce the infant death whereas it does
not matter much in case of fathers. Thus, the recent goal for providing 1‐8 years of schooling
through Right to Education (RTE) Act 2009 in India should specifically focus on reducing the
gender inequality in education by targeting the women for reducing the infant death, in
addition to the other objectives specified in the act.
The rural‐urban differential of infant mortality has been well documented in the context of
many developed and developing countries. A few research studies on infant and child
mortality in India have shown that the risk of infant death is relatively low in urban areas as
compared to rural areas (Gupta 1990; NIMS et al, 2012; Saikia et al, 2013). Narrating the
reasons, studies have found that parent’s education plays an important role. Illiterate parents
in the rural region go for early marriage of their children (particularly daughters) which
further leads to high risk of infant death. Very young mothers (particularly under the age of
20) may experience difficult pregnancies and deliveries because of their physical immaturity.
They are also likely to have limited knowledge and confidence in caring for infants and
young children (Pandey et al, 1998). The study by Babson and Clarke (1983) have found that
in the neonatal period, a significantly greater overall mortality occurred in infants born to the
younger mothers and the reason was to give birth of low weight babies.
As shown in Table 2, for major states of India, EAG and Non‐EAG states, the IMR in rural
region is higher than that of urban region. But, surprisingly, the rural‐urban gap in the IMR is
higher for Non‐EAG states (18.3) compared to EAG states, which is 9.9 points. Usually, it is
expected that the rural‐urban difference in IMR is less for the states with higher socio‐
economic developments. This may be explained by the fact that the Non‐EAG states have
given due focus on urban development neglecting rural region, which is an ongoing policy
debate in most of the developing countries, including India. Though, in both rural and urban
regions of EAG and Non‐EAG states, the IMR declines with the increase in mother’s
educational attainment, the pattern of reduction provides some interesting picture (Table 2).
First, the IMR among the children born to illiterate mothers is 3.1 and 1.7 times higher than
those born to mothers with post secondary level of education in urban and rural region
respectively in EAG states. But, for Non‐EAG states it is 2 times higher both in urban and
rural region. This reveals the extent of rural‐urban gap in the IMR by mother’s educational
attainment between EAG and Non‐EAG states. More clearly, the change in the levels of IMR
by mother’s education is more or less same in rural area in both EAG and Non‐EAG states,
whereas the patterns of change in IMR differ widely in the urban region. Table 2 shows that in
the urban region of EAG states the IMR declines by 16.6 points with the increase in the
15
Table 2
Infant mortality rates by parents education and region
in EAG and Non EAG states of India
Variables EAG states Non‐EAG states India
Rural Urban Rural Urban Rural Urban
Mother Education
No education 73.0 81.1 58.5 40.4 69.2 64.1
1‐8 years complete 66.3 64.5 48.5 41.4 56.8 49.2
9‐10 years complete 55.1 42.4 40.5 25.5 47.3 30.9
> 10 years complete 41.5 25.8 28.7 19.6 33.2 21.7
Father Education
No education 73.6 77.6 55.9 35.8 67.8 60.0
1‐8 years complete 70.3 75.6 55.3 31.8 65.8 55.7
9‐10 years complete 61.1 55.2 45.2 30.8 55.2 42.8
> 10 years complete 53.1 32.9 33.7 24.3 44.8 27.3
Parents Education
Both are Illiterate 73.1 86.2 52.0 35.6 67.2 66.1
Both have completed > 10 years 37.8 27.7 17.4 22.5 26.1 24.5
Total 69.1 59.2 49.5 31.2 61.7 43.0
Source: Author’s calculation from NFHS‐3 data.
mother’s education form illiterate to elementary, whereas in urban region of Non‐EAG states
there is no change in the IMR among children born to illiterate mothers and mothers with
elementary level of education and further studies are needed to see the reasons of this.
Like mother’s education, increase in the father’s educational level also reduces the IMR both
in rural and urban region in EAG states, Non‐EAG states and major states of India (Table 2).
In all the cases, the reduction is minimum with the increase in the father’s educational level
from illiterate to elementary level of education. The fall of IMR is more or less same with the
further increase in the educational level of the father (i.e. from 1‐8 years of schooling to 9‐10
years of schooling and then to more than 10 years of schooling) in all the cases, except urban
region of Non‐EAG states, where the reduction in IMR is only one point between 1‐8 years
and 9‐10 years of schooling. Thus, as Table 2 shows, the attainment of secondary and above
level of education by fathers matter more in the reduction of IMR than attaining compulsory
level of schooling. More clearly, there is not much difference in IMR level among children
born to illiterate fathers and father having elementary level of education in the regional level
in India.
To relate the parent’s education (father and mother education taken together) with the infant
mortality rate in the regional level, their level of education were combined and categorised as
(a) both are illiterate, and (b) both have senior secondary and above level of education and is
presented in Figure 1. The cases where either of the parents is illiterate and both of them have
some level of education (i.e. less than 11 years of schooling) are excluded from this particular
analysis. The intention in this case was to find out the difference in the IMR with two extreme
16
Figure 1
Infant Mortality Rate by Parents Education and Region
level of education of the parents, i.e., both of them are illiterate and both of them have more
than 10 years of schooling by region. As expected, the IMR has reduced by 24.7 points, 15.7
points and 20.1 points in EAG states, Non‐EAG states and major states of India respectively
when parents educational level changes from ‘illiterate’ to ‘both have senior secondary and
above level of education’ as shown in Fig. 1. In comparison to the individual level of
education of the father and mother, their combined education reduces the IMR to a greater
extent. This is in conjunction with a study from Africa by Kiros and Hogan (2001) that reports
a significant reduction of children death with the parents having any level of education
compared to when both parents are illiterate.
4.2 Role of Parental Education in the Regional variation of IMR: Logit Estimates
We now turn to analyse the role of parental education in the regional variation of infant
mortally using logit models. To check the sensitivity of the results four sets of regressions are
estimated. The first three models examine the relationship between infant death and parental
education without controlling for other variables, whereas the fourth model analyses the
impact of parent’s education by controlling the potential confounding factors. In each case,
five different equations are estimated to examine the relative risk of infant death by regions.
These models include one each for EAG, Non‐EAG and major states of India; and within
major states of India, two separate equations are estimated to examine the rural‐urban
differences. The equations have considered ‘whether the infant has died or not’ as dependent
variable and parent’s education (mother and father separately), mother’s exposure to mass
media, socio‐economic empowerment of the mother, other related socio‐economic factors,
accessibility to health care, and demographic characteristics as explanatory variables. The
74.4
49
67
31.8
21.3 25.1
0
20
40
60
80
EAG States Non‐EAG States India
Infant Mortality rate
Parents are Illiterate Parents have > 10 years of Education
17
notation and definition of the variables are given in Table A1 in appendix. The logit results are
presented in Table 3 & 4 and are discussed under the following headings.
Table 3
Logit Estimates of Infant Death by Parental Education (Model 1 to 3)
Variable EAG States
(Odds Ratio)
Non‐EAG States
(Odds Ratio)
India
(Odds Ratio)
Rural Urban Total
Model 1
Moth_elem 0.88*
(0.06)
0.88
(0.08)
0.87*
(0.08)
0.81***
(0.05)
0.81***
(0.04)
Moth_sec 0.68***
(0.06)
0.62***
(0.06)
0.57***
(0.06)
0.63***
(0.05)
0.58***
(0.04)
Moth_higher 0.38***
(0.05)
0.40***
(0.05)
0.35***
(0.04)
0.43***
(0.06)
0.35***
(0.03)
Log‐Likelihood ‐5413.0 ‐3563.6 ‐2900.6 ‐6108.2 ‐9017.7
Constant ‐2.53 ‐2.93 ‐2.78 ‐2.62 ‐2.65
Observations 22778 21821 17152 27447 44599
Model 2
Fath_elem 1.04
(0.07)
0.86*
(0.08)
0.78**
(0.09)
1.01
(0.07)
0.92
(0.05)
Fath_sec 0.78***
(0.06)
0.75
(0.07)
0.74***
(0.08)
0.74***
(0.05)
0.72***
(0.04)
Fath_higher 0.57***
(0.05)
0.48***
(0.05)
0.46***
(0.05)
0.59***
(0.05)
0.50***
(0.03)
Log‐Likelihood ‐5332.6 ‐3560.8 ‐2894.4 ‐6039.1 ‐8954.1
Constant ‐2.53 ‐2.92 ‐2.79 ‐2.63 ‐2.66
Observations 22449 21672 17000 27121 44121
Model 3
Moth_elem
0.91*
(0.07)
0.91
(0.08)
0.89*
(0.09)
0.83***
(0.06)
0.83***
(0.05)
Moth_sec 0.77***
(0.07)
0.66***
(0.07)
0.58***
(0.07)
0.70***
(0.06)
0.63***
(0.04)
Moth_higher 0.45***
(0.07)
0.47***
(0.07)
0.37***
(0.06)
0.50***
(0.08)
0.40***
(0.04)
Fath_elem 1.07
(0.08)
0.94
(0.09)
0.86
(0.10)
1.08
(0.07)
1.00
(0.06)
Fath_sec 0.86**
(0.06)
0.93
(0.10)
0.98
(0.11)
0.86**
(0.06)
0.90*
(0.05)
Fath_higher 0.82*
(0.09)
0.76**
(0.11)
0.89*
(0.13)
0.80**
(0.08)
0.83**
(0.07)
Log‐Likelihood ‐5317.2 ‐3544.7 ‐2869.2 ‐6023.2 ‐8904.6
Constant ‐2.51 ‐2.88 ‐2.73 ‐2.59 ‐2.62
Observations 22447 21672 16998 27121 44119
Notes: (a) * P 0.10, ** P 0.05, *** P 0.01 (b) The entries in parenthesis refer to standard error. Source: Author’s self‐computation.
18
Table 4
Logit Estimates of Infant Death by Parental Education (Model 4)
Variable EAG States
(Odds Ratio)
Non‐EAG States
(Odds Ratio)
India
(Odds Ratio)
Rural Urban Total
Moth_elem
0.76***
(0.08)
0.94
(0.12)
0.78***
(0.07)
0.83*
(0.11)
0.80***
(0.06)
Moth_sec 0.62***
(0.08)
0.65***
(0.10)
0.64***
(0.08)
0.55***
(0.09)
0.59***
(0.06)
Moth_higher 0.37***
(0.08)
0.49***
(0.11)
0.51***
(0.12)
0.36***
(0.08)
0.40***
(0.06)
Fath_elem 1.05
(0.10)
0.99
(0.13)
1.08
(0.09)
0.90
(0.14)
1.03
(0.08)
Fath_sec 0.95
(0.09)
0.83*
(0.17)
0.90*
(0.10)
0.82*
(0.14)
0.85*
(0.08)
Fath_higher 0.73*
(0.14)
0.64*
(0.19)
0.72*
(0.15)
0.65*
(0.17)
0.67*
(0.12)
Moth_medexp 0.81**
(0.08)
0.92*
(0.11)
0.70*
(0.08)
0.92*
(0.18)
0.81*
(0.08)
Moth_empw 0.78***
(0.06)
1.00
(0.10)
0.85**
(0.06)
0.84***
(0.08)
0.83***
(0.05)
Middle_income 1.00
(0.10)
0.86*
(0.12)
0.93*
(0.09)
0.92
(0.17)
0.92*
(0.07)
Rich 0.98
(0.13)
0.78*
(0.13)
0.77**
(0.10)
0.95*
(0.15)
0.90*
(0.09)
Moth_working 0.90*
(0.06)
0.93*
(0.08)
0.86**
(0.06)
0.98*
(0.08)
0.90**
(0.05)
Caste 0.92*
(0.07)
0.96*
(0.11)
0.98*
(0.07)
0.95
(0.10)
0.97*
(0.06)
Dwater_pipe 0.90
(0.09)
0.96
(0.10)
0.84**
(0.08)
0.78**
(0.08)
0.80**
(0.05)
Toilet_Facility 0.94
(0.04)
0.82*
(0.10)
0.80**
(0.08)
0.98*
(0.15)
0.90*
(0.07)
Delivery_place 0.92**
(0.11)
0.77**
(0.11)
0.85**
(0.10)
0.83**
(0.10)
0.82**
(0.07)
Moth_20_29 0.66***
(0.11)
0.76*
(0.17)
0.67***
(0.10)
0.88
(0.28)
0.70***
(0.09)
Moth_30_39 0.51***
(0.10)
0.46***
(0.12)
0.45***
(0.08)
0.70*
(0.24)
0.49***
(0.08)
Moth_40_49 0.39***
(0.11)
0.82
(0.32)
0.55***
(0.13)
0.20**
(0.13)
0.47***
(0.11)
Birth_Order_2 0.75***
(0.08)
0.76**
(0.09)
0.73***
(0.07)
0.77**
(0.10)
0.74***
(0.06)
Birth_Order_3 0.74***
(0.09)
0.84*
(0.13)
0.74***
(0.08)
0.90
(0.15)
0.79***
(0.07)
19
Variable EAG States
(Odds Ratio)
Non‐EAG States
(Odds Ratio)
India
(Odds Ratio)
Rural Urban Total
Birth_Order_4 0.96
(0.12)
1.08
(0.21)
1.03
(0.13)
1.03
(0.21)
1.03*
(0.11)
Birth_Order5 1.39***
(0.18)
1.67***
(0.35)
1.42***
(0.18)
2.09***
(0.42)
1.55***
(0.17)
HH_size 0.91***
(0.01)
0.90***
(0.02)
0.93***
(0.01)
0.89***
(0.02)
0.92***
(0.01)
Log‐Likelihood ‐3355.1 ‐2033.9 ‐2624.8 ‐1782.3 ‐5429.3
Constant ‐1.34 ‐1.62 ‐1.43 ‐1.83 ‐1.46
Observations 19824 18429 23118 15135 38253
Notes: (a) * P 0.10, ** P 0.05, *** P 0.01 (b) The entries in parenthesis refer to standard error. Source: Author’s self‐computation.
4.2.1 Effect of Mother’s Education
The strong association between mothers’ education and infant/child mortality is considered
as one of the most consistent and powerful finding in public health (Caldwell, 1979; Bicego
1993; Ozaltin et al, 2010). A large body of research suggests that a casual relationship exists
between maternal education and childhood health and mortality. In the past 30 years, many
hypotheses have been proposed for the mechanisms through which increased education
could lead to reductions in child mortality, including individual effects through improved
use of health services, economic advantages, empowerment and independence of women,
and community level effects (Gakidou et al, 2010). Caldwell (1979) suggested that educated
mothers are more likely to shift from a ‘fatalistic’ acceptance of health outcomes towards the
implementation of simple‐health promoting practices. This often includes an increased
capacity to manipulate modern medical systems.
Consistent with other studies, the results for major states of India as reported in Table 3 (model
1 and model 3) shows that, mother’s education has a significant effect on reducing the
probability of infant death. Children of mothers who attended compulsory level of schooling
i.e. elementary are less likely to die than are children of mothers who are illiterate. Children of
mothers with senior‐secondary and above level of education are the least likely to experience
infant deaths. Similar finding is noticed in several studies undertaken both in India and
international context (Desai and Alva, 1998; Mondal et al, 2009; Adeyele and Ofoegbu, 2013;
Saikia et al, 2013). In EAG and Non‐EAG states and also in rural and urban regions the
relative risk of infant mortality among the children born to educated mothers is considerably
less as compared to the illiterate mothers, except mother’s elementary education in Non‐EAG
states, which is statistically not significant. Furthermore, as expected, the probability of infant
death declines with increasing the levels of education of the mothers. More clearly, children
born to the mothers having more than 10 years of education (considered as the highest level
of education in the analysis) have least risk of dying than the children born to the mothers
having lower levels of education i.e. mothers with the educational level of elementary and
20
secondary, in all the regions (rural‐urban and EAG‐Non‐EAG states). Also, the effect of
mother’s education on the probability of infant death shows the expected results when taken
together with father’s education (Model 3). It is usually expected that mother’s education is
correlated with father’s education and hence taken together may altered the results, which is
not the case in the present analysis. Even after controlling for the cofounding factors, the
effects of mother’s education in reducing number of infant deaths remained consistent and
strongly significant (Model 4). All four regression models suggested a significant reduction in
the infant death in the regional level in India with the increase in the mother’s education. This
strongly calls for educating mothers to reduce the infant death in India, irrespective of
regions.
Logit estimates show that attending compulsory schooling (elementary level of education) by
the mother reduces the infant death in EAG states whereas it does not matter in Non‐EAG
states, as it is statistically not significant (Model 1, 3 and 4). However, the relative risk of infant
mortality among the children born to the mothers having 9 to10 years of schooling is lower in
Non‐EAG states as compared to EAG states. The difference in the odds ratio is highest in
Model 3 i.e., 11 percentage points. But the secondary and above level of education of the
mother matters more in EAG states than Non‐EAG states in reducing the infant death. Thus,
attaining above compulsory level of education by the mother matters more in reducing the
probability in infant death in Non‐EAG states, whereas in EAG‐states, all levels of education
matters. This may be due to the fact that in relatively developed states, there may not be
much difference between the mothers having no education and having compulsory level of
education on reducing infant death.
Studying the role of parental education (particularly mother’s education) on infant mortality
in rural and urban regions is extremely important in Indian context. It is because of the fact
that 1) the infant mortality rate in rural India is significantly higher than urban India and 2)
the female literacy rate is lower in rural areas compared to urban regions. According to the
2011 census of India, the IMR is 66 is urban region and 38 in rural region. Similarly, the
female literacy is 79.1 in urban areas and 57.9 in rural areas. The attempt to find the effect of
mother’s education on infant death by region may address the concern that the low level
female education is a major hindrance for the rural India to reduce its infant death to that of
urban India. This is found to be true in the present analysis. The logit estimates in Model 4
show that for all level of education of the mother the probability of infant death is lower in
urban region as compared to rural region. The difference in the odds ratio being highest (30
percentage points) in case of mothers attainted above secondary level of education. However,
the results in Model 1 and 3 gives mixed results.
4.2.2 Effect of Father’s Education
As we have already discussed, several studies have attempted to estimate the impact of
mother’s education on child health and mortality related issues, including the infant death.
However, the impact of father’s education on these aspects is less explored. This is mainly
21
because of the established hypothesis that the mothers play a pivotal role in taking care of the
babies in their early days of birth, which of course decides the child health and mortality
status to greater extent. But in the changing context, the role of fathers in taking care of their
babies, even in the early days of their birth is a reality, though this is more seen in urban
region than that of rural region. Thus, one can expect the role of father’s education and
awareness on his child’s health status and mortality. This is found to be true to some extend
in the present analysis. The estimated regression results presented in Table 4 show that the
probability of infant death is less for the children whose fathers have secondary, senior
secondary and higher level of education compared to the children whose fathers are illiterate.
In model 3 and 4 attending elementary level of education by the father does not matter in
reducing the risk of infant death as the coefficients are statistically not significant. However,
in model 2, the impact of elementary education of the father on infant death is negative and
statistically significant in Non‐EAG states and rural region. Similarly, for major states of
India, infants of the father having elementary level of education experience lower risk of
dying than the infants of illiterate fathers (although this result did not reach statistical
significance). The probability of infant death is lower for the children whose fathers have
secondary level of education as compared to the fathers having elementary level of
education. Similarly, the infant death is found to be less among the children whose fathers’
having senior secondary and above level of education than having secondary level of
education. Attaining senior secondary and above level of education by the father significantly
reduces the probability of infant death across all the regions and major states of India. These
results suggests for providing minimum of secondary and above level of education to the
fathers for an effective reduction of Infant mortality in India.
Comparing the results between EAG and Non‐EAG states, it is found that the effect of the
highest level of education of the father (i.e. senior secondary and above level of education) on
infant death is found to be significantly higher among Non‐EAG states than EAG states, the
difference in the odds ratio being 9 percentage points after controlling the confounding
factors (Model 4). The odds ratio of the secondary level of education of the father (Fath_sec) is
statistically significant in EAG states in Model 2 and 3 but statistically not significant in Model
4. Thus, though all level of education of the father matters in Non‐EAG states, attaining
compulsory and above senior secondary level of education matters in EAG states in reducing
infant death. As expected, the effect of father’s education on reducing infant death is higher in
urban region compared to rural region when controlled for other variables. But the effect of
secondary level of education of the father on the probability of infant death is statistically not
significant in urban region.
4.2.3 Effect of Mother’s Mass Media Exposure and Socio‐Economic Empowerment
It is expected that controlling all other factors, a mother’s exposure to different mass media
may reduce the mortality of her children. It may be due the fact that women who are exposed
to mass media such as television, radio and newspaper are likely to have access to
information on health‐care services and ways of enhancing maternal and child health.
22
Mother’s exposure to mass media may also act as an indicator of the economic status of the
household. In this analysis, a woman is considered to be exposed to mass media if she listens
to radio or watches television or read newspaper at least once a week. Logit estimates show
that the children born to the mothers having any kind of exposure to the mass media have
lower probability of death in their infant stage compared to the children born to the mothers
having no mass media exposure. It is found to be true in all the five equations estimated in
the present analysis and the coefficients are statistically significant at 10 per cent level of
significance. Interestingly, the effect of mass media exposure on the probability of infant
mortality is higher in EAG states than Non‐EAG states, the odds ratio being 0.81 and 0.86
respectively. Similarly, mother’s mass media exposure considerably reduces the probability
of infant death in rural region as compared to urban region. Thus, the results strongly suggest
that mother’s engagement with the mass media such as watching television, listening radio
and reading news paper is an important factor in reducing infant mortality across the
regions. This works more effectively in the regions of lower socio‐economic status of the
people like EAG states and in rural region.
Another important factor which is statistically significant in the determination of infant
mortality in EAG and major states and also in rural and urban region is the socio‐economic
empowerment of the mother. Surprisingly, it turned out to be statistically not significant in
Non‐EAG states. The result obtained from the analysis may be interpreted as follows: the
relative risk of infant mortality is relatively lower if mother alone take decision on different
socio‐economic aspects of the family (such as mother’s health care expenditure, decision on
large and daily household purchases, decision on spending husband’s money, and decision
on visit to family and relatives) as compared to the mother alone does not take such
decisions. Similar finding was revealed in a study by Adeyele and Ofoegbu (2013) in
Nigerian context. Thus, mothers should be allowed to take decisions on both social and
economic matters of the family which will lower down the risk of infant death. This clearly
supports the fact that education gives women the power and the confidence to take decision‐
making into their own hands which helps further to reduce the infant mortality. It is
important to note that in EAG states, the effect is higher (than Non‐EAG states) and
statistically significant at 1 per cent level of significance. Between rural and urban region, the
effect is higher in the latter one which supports the general expectation.
4.2.4 Effect of Other Controlling Factors
The other controlling factors included in the logit estimation can broadly be categorised as
socio‐economic (economic status of the household measured in terms of wealth index,
mother’s working status, and caste), sanitation (in terms of availability of toilet facility),
sources of drinking water, place of delivery, demographic factors (mother’s age, birth order
of the child, and size of the household).
Among the socio‐economic factors, the impact of working status of the mother on the
probability of infant death gives some interesting findings. The children born to the working
23
mothers have lower probability of dying in the infant stage as compared to the children born
to the mothers who are not engaged in the job market. Contrary to this, the study by Mustafa
and Odimegwu (2008) shows that the risk of infant death is more for the children born to
mother working in the agriculture compared to those whose mothers are not working. The
finding of the present study does not go with the general assumption that the mothers who
are not working in the job market provides better care to their new born children, which
reduces the probability of their death in the early ages. The finding rather supports the fact
that working mothers have better socio‐economic status and also aware of better health care
facilities that keeps their children in good health. It is interesting to note that the effect of the
mother’s working condition on reducing the probability of infant death is higher in EAG
states and rural areas compared to Non‐EAG states and urban region. This may be due the
fact that in these regions the difference in the awareness and other related aspects differs
considerably between mothers engaged in the labour force and not engaged in the labour
force. Thus, this suggests for encouraging the women to engage in the labour force
(particularly in rural region and EAG states) to reduce the infant mortality. As expected, the
logit results show that with the increase in the economic status of the households the
probability of infant death declines. However, it is surprising to note that it is statistically not
significant in EAG states and urban region. The social category/caste of the mother gives the
expected results. In all the five estimated equations, children born to scheduled caste and
scheduled tribe households have higher probability of dying in the infant stage compared to
the children born to forward caste and other backward classes. The effect of caste in the urban
region is found to be statistically not significant.
Access to toilet (particularly, a flush or pit toilet) is potentially a very important determinant
of infant and child mortality in developing countries, including India. Children in the
households that lack such access could have higher probability of affecting with diseases
tetanus and digestive disorders than other children (Puffer and Serrano 1973; United Nations,
1985). Quite a few studies have examined the impact of the availability and in some cases the
quality of toilet facility on the probability of infant and child mortality, both in India and
international context and have found that the access to toilet facilities in household
considerably reduces the infant and child mortality (Pandey et al, 1998; Kembo and Ginneken,
2009; Mondal et al, 2009). The logit estimates in the present analysis reveals the similar
finding. However, the impact of the availability of toilet facility in EAG states is found to be
statistically not significant. The impact of toilet facility towards reducing the infant mortality
in rural region is higher than urban region. The value of odds ratio in rural region is 0.77
whereas it is 0.98 in urban region. This convey the message that the providing toilet facility in
rural areas may work as an effective panacea to reduce the infant mortality. It is important to
note that the proportion of households without any toilet facility is much greater in rural
areas (74%) than in urban areas (17%) in 2005‐06.
The effect of the source of drinking water—used as a dummy variable (takes the value 1, if
piped water and 0, if from other sources like tube well, drug well and surface) shows the
expected result in all the five models. Children born to the households having access to pipe
24
drinking water are less likely to die in their infant stage compared to the children born to the
households getting drinking water from other sources. Quite a few studies in developing and
underdeveloped countries have revealed the similar findings (Hossain and Islam, 2008;
Kembo and Ginneken, 2009). The effect of the sources of drinking water in reducing the
probability of infant death is higher in urban region than that of rural region. However, the
coefficients of the sources of drinking water in EAG and non‐EAG states are found to be
statistically not significant.
It is evident from the result that (model 4, Table 4) the probability of infant death is lower
among the mothers who have gone for institutional deliveries as compared to the mothers
who have delivered their babies in the home. The value of the odds ratio is 0.82 for major
states of India. Similarly, there is also an inverse relationship between the place of delivery
and the risk of infant death in the regional level. However, as expected, the effect is higher in
urban region than that of rural region. It may be due to the fact that the existing hospitals in
the rural region are lacking the trained medical professionals than that of urban region. Thus,
the new born babies in the rural hospitals are not getting proper care which leads for infant
deaths. Also, the effect of institutional delivery on reducing the risk of infant death is higher
in Non‐EAG states as compared to EAG states. The result may imply the availability of better
quality health care services in the Non‐EAG states in India compared to EAG states.
Demographic characteristics may have an effect on the probability of infant death. In the
present analysis four different demographic factors are considered in the logit estimation.
These are age of the mother, birth order of the child, and size of the household. The
supporting arguments for including these variables in the estimation are given in section 3 on
data and methodology. Logit estimates clearly demonstrate that the probability of infant
death is lower for the children born to the older mothers (age more than 20) as compared to
younger mothers, age being less than 20. This is true for all the five equations estimated in
this study. But the coefficient for the urban region is statistically not significant, which may be
due to the considerably lees sample size in this group i.e., less than 2 per cent. The relative
risk of infant mortality declines further with the increase in the mother’s age from 20‐29 to 30‐
39 and the coefficients are statistically significant at 1 per cent level, except the urban region
which is significant at 10 per cent level of significance. Increase in the mother’s age from 30‐39
to 40‐49 gives mixed results for different regions of India. In case of major states, and rural
region the risk of infant mortality increases whereas in urban region and EAG states it
declines. The coefficient associated with Non‐EAG states is statistically not significant. In each
case the effect of mother’s age on infant mortality is found to be higher in rural region and
EAG states as compared to urban region and non‐EAG states respectively. The results
analysed here partially supports the general argument that the children born to mothers
under 20 or around 40 years old are likely to have elevated risks of mortality. Very young
mothers may experience difficult pregnancies and deliveries because of their physical
immaturity. They are also likely to have limited knowledge and confidence in caring for
infants and young children. Women around 40 years old may also experience age‐related
problems during pregnancy and delivery (Pandey et al, 1998). A similar picture emerges in a
25
study by Babson and Clarke (1983). They have found that in the neonatal period, a
significantly greater overall mortality occurred in infants born to the younger mothers and
the reason was to give birth of low weight babies by younger mothers.
As already noted in section 3, the relation between birth order and infant mortality is not well
understood. But it is generally expected that higher birth order (to a certain extent) reduces
the mortality risk. The logit results show that the risk of mortality is less for the second and
third born child as compared to first born child and it is true in all the five equations.
However, the probability of infant death increases with the further increase in the birth order.
This confirms the U shape of the relationship between birth order of the child and infant
death. The studies in the context of under developed countries like Zimbabwe and Kenya
shows similar findings (Mustafa and Odimegwu, 2008; Kembo and Ginneken, 2009). Infants
of birth order six or higher are 2.75 times more likely to die in infancy relative to births of
order two through five.
The size of the household may have a direct or inverse relationship with the infant mortality.
The results for all the five cases here show that the children born in the large households are
less likely to die in their infant stage than the children born to small families. This may be due
to the fact that the children born to joint families might have got good health care from their
elders which help them to be well. On the contrary, children born to small households
(usually nuclear families) may not get good health care from other family members which
increase their risk of dying in the early stage of their birth. However, this needs to be
examined in detail. The logit results show that in the major states of India the probability of
female infant mortality is higher than male children and the coefficient is significant at 10 per
cent level of significance.
5. Conclusions
The present study has primarily examined the role of mother’s and father’s education in
reducing the infant mortality in EAG‐Non‐EAG states and rural‐urban regions in India.
Although the main focus of the paper was to examine the effect of education, an attempt was
also made to analyse the impact of mother’s exposure to mass media and her socio‐economic
empowerment that are closely related to education on infant mortality. Besides these, the
effect of other confounding factors broadly be categorised as socio‐economic (economic status
of the household measured in terms of wealth index, mother’s working status, and caste),
sanitation (in terms of availability of toilet facility), sources of drinking water, accessibility to
health care, and demographic factors (mother’s age, birth order of the child, and size of the
household,) on infant mortality is also examined. The paper has uncovered several
interesting findings—some expected and some unexpected.
While the overall infant mortality is 57 in major states of India, the analysis shows that it
varies enormously by parental education and regions. As expected, the IMR is higher in rural
region and EAG states compared to urban region and Non‐EAG states in India. Educational
26
attainment of both mother and father has an inverse relationship with the IMR in major states
of India, EAG ‐Non‐EAG states, and between rural‐urban regions. However, there exists a
wide regional variation in the rate of decline in the IMR by level of education of the mother
and father. The intention to examine the difference in the IMR with two extreme level of
education of the parents, i.e., both of them are illiterate and both of them have more than 10
years of schooling shows the expected result and suggests that in comparison to the
individual level of education of the father and mother, their combined education reduces the
IMR to a greater extent.
The regression results show that both mother’s and father’s education are significantly
associated in reducing the infant mortality across the regions and major states of India,
although the relative effect of different levels of education of the parents varies between
EAG‐Non‐EAG states and rural‐urban regions. Analysis suggests that attaining above
compulsory level of education by the mother matters in reducing the probability in infant
death in Non‐EAG states, whereas in EAG‐states, all levels of education matters. This may be
due to the fact that in relatively developed states, there may not be much difference between
the mothers having no education and having compulsory level of education on reducing
infant death.
As per the father’s education is concerned, in most of the cases the effect of elementary
education on the risk of infant death is found to be statistically not significant. However, the
attaining secondary, senior secondary and higher level education gives the expected results.
Thus, it calls for providing minimum of secondary and above level of education to the fathers
for an effective reduction of Infant mortality in India. But this result should not be construed
as implying that taking up to elementary level of education by the fathers is not important.
Rather, it suggests that initiatives should be taken to educate fathers beyond elementary
which has a strong effect on reducing infant mortality in India. This study was able to find
out that the children born to the mothers having any kind of exposure to the mass media
have lower probability of death in their infant stage compared to the children born to the
mothers having no mass media exposure and it works more effectively in the regions that are
underdeveloped such as EAG states and rural areas. The effect of different controlling factors
included in the logit model gives more or less expected results.
The paper concludes that the benefits of parental education in reducing infant death endure
even when other socio‐economic factors are taken into account. The findings reached in this
paper emphasise to provide education to the parents, particularly to mothers of backward
regions or states, which are crucial for reducing infant mortality. Besides parent’s education
attempt should also be made to increase the scope of getting mass media exposure and
higher level of socioeconomic empowerment of the mother (two factors that are closely
related with education) to reduce the infant mortality in India.
27
References
Adeyele, I.T. and D.I. Ofoegbu (2013), Infant and Child Mortality in Nigeria: An Impact
Analysis, International Journal of Economic Practices and Theories, 3 (2): 122‐32.
Babson, S.G., and N.G. Clarke (1983), Relationship between infant death and maternal age.
Comparison of sudden infant death incidence with other causes of infant mortality,
Journal of Pediatrics, 103 (3): 391‐93.
Basu, A. M. and R. Stephenson (2005), Low Levels of Maternal Education and the Proximate
Determinants of Childhood Mortality: A Little Learning is Not a Dangerous Thing,
Social Science and Medicine, 60 (9): 2011‐23.
Bobak, M., H. Pikhart and I. Koupilová (2000), Maternal socioeconomic characteristics and
infant mortality from injuries in the Czech Republic 1989–92, Injury Prevention; 6
(3):195‐98
Census of India (2011), Primary Census Abstract, Figures at a Glance, New Delhi: Office of
the Registrar General of India.
Caldwell, J. (1979), Education as a Factor in Mortality Decline: An Examination of Nigerian
Data, Population Studies, 33 (3): 395‐413.
Caldwell, J. (1994), How is greater maternal education translated into lower child mortality?
Health Transition Review, 4 (2): 224‐29.
Caldwell, J. and McDonald, P. (1982), Influence of maternal education on infant and child
mortality: levels and causes, Health Policy and Education, 2 (3‐4): 251‐67.
Das, N.P. and D. Dey (2003), Understanding the causative factors behind stalling of infant
mortality in India during the recent period, Demography India, 32 (2): 249‐73.
Desai, S. and Soumya Alava (1998), Maternal Education and Child Health: Is there a Strong
Causal Relationship? Demography, 35 (1): 71‐81.
Gakidou, E., K. Cowling, R. Lozano, and C.J.L. Murray (2010), Increased educational
attainment and its effect on child mortality in 175 countries between 1970 and 2009: a
systematic analysis, Lancet, 376 (9745): 959‐74.
Gemperli, A., P. Vounatsou, I. Kleinschmidt, M. Bagayoko, C. Lengeler, and T. Smith (2004),
Spatial Patterns of Infant Mortality in Mali: The Effect of Malaria Endemicity,
American Journal of Epidemiology, 159 (1): 64–72.
Govindasamy, P. and B.M. Ramesh (1997), Maternal education and utilization of maternal
and child health services in India, NFHS Subject Reports, Number 5, IIPS, Mumbai.
Gunasekaran, S. (2005), Differentials and Determinants of Infant and Child Mortality in Rural
India, XXV International Population Conference, Tours, France, July 18‐23,
http://iussp2005.princeton.edu/papers/51367 (Accessed on 5th December 2012).
Gupta, M.D. (1990), Death Clustering, Mothers’ Education and the Determinants of Child
Mortality in Rural Punjab, India, Population Studies, 44 (3): 489‐505.
Hossain, M. and M. Islam (2008), Effects of Demographic and Household Variables on Infant
and Child Under‐five Mortality: An Application of Logistic Model, The Internet
Journal of Health, 8 (2): 1‐9.
International Institute for Population Sciences (IIPS) ORC Macro (2007), National Family
Health Survey (NFHS‐3), 2005‐06: India. Vol. 1, Mumbai: IIPS.
28
Kapoor, S. (2010), Infant mortality rates in India: District level variations and correlations,
http://www.isid.ac.in/~pu/conference/dec_10_conf/Papers/ShrutiKapoor.pdf (Accessed
on 5th December 2012)
Kembo, J. and J.K.V. Ginneken (2009), Determinants of infant and child mortality in
Zimbabwe: Results of multivariate hazard analysis, Demographic Research 21 (13):
367‐84.
Khasakhala, A.A. (2003), Effect of maternal education on infant survival in rural Kenya,
Demography India, 30 (2): 93‐106.
Kiros, G.E. and D.P. Hogan (2001), War, famine and excess child mortality in Africa: the role
of parental education, International Journal of Epidemiology, 30 (3): 4447‐55.
Kravdal, Q. (2004), Child mortality in India: The community‐level effect of education,
Population Studies, 58 (2): 177‐92.
NIMS, ICMR and UNICEF (2012), Infant and Child Mortality in India: Levels, Trends and
Determinants, National Institute of Medical Statistics, Indian Council of Medical
research, and UNICEF India Country Office, New Delhi, India.
Pandey, A., M.K. Choe, N.Y. Luther, D. Sahu, and J. Chand (1998), Infant and Child Mortality
in India, National Family Health Survey Subject Reports, Number 11, IIPS, Mumbai.
Papageorgiou, C. and P. Stoytcheva (2012), Education Inequality among Women and Infant
Mortality: A cross‐country empirical investigation, Progress in Economic Research,
Nova Science Publishers, New York, NY.
http://www.chrispapageorgiou.com/papers/PS.pdf (Accessed on 5th November 2012)
Puffer, R.R., and Carlos V. Serrano (1973), Patterns of mortality in childhood: Report of the
Inter‐American Investigation of Mortality in Childhood. Washington, D.C.: Pan
American Health Organization.
Rama Rao S., A. Pandey, and K.I. Shajy (1997), Child Mortality in Goa: A cross‐sectional
analysis, Biodemography and Social Biology, 44 (1‐2): 101‐10.
Registrar General of India (2011), Sample Registration System Statistical Report, Vital
Statistics Division, New Delhi: Office of the Registrar General of India.
Singh, A., P.K. Pathak, R.K. Chauhan, and W. Pan (2011), Infant and Child Mortality in India
in the Last Two Decades: A Geospatial Analysis, PLoS ONE, 6 (11): 1‐19.
Saikia, N., A. Singh, D. Jasilionis, and F. Ram (2013), Explaining the rural‐urban gap in infant
mortality in India, Demographic Research, 29 (18): 473‐506.
Singh‐Manoux, A., A. Dugravot, G. D. Smith, M. Subramanyam, and S.V. Subramanian
(2008), Adult education and child mortality in India: the influence of caste, household
wealth, and urbanization, Epidemiology, 19 (2): 294‐301.
Titaley, C., M.J. Dibley, K. Agho, C.L. Roberts, J. Hall (2008), Determinants of neonatal
mortality in Indonesia. BMC Public Health 8:232.
Uddin, J., and Z. Hossain (2008), Predictors of Infant Mortality in a Developing Country,
Asian Journal of Epidemiology, 1 (1):1‐16.
United Nations (1985), Socio‐economic differentials in child mortality in developing
countries, Department of International Economic and Social Affairs, New York: United
Nations.
29
Wang, L. (2003), Determinants of child mortality in LDCs: Empirical findings from
demographic and health surveys, Health Policy, 65 (3): 277‐99.
Ware, H. (1984), Effects of Maternal Education, Womenʹs Roles, and Child Care on Child
Mortality, Population and Development Review, 10 (supplement): 191‐214.
World Health Organization (2012), World Health Statistics, Geneva: WHO Press.
30
Appendix
Table A1
Notation and Definition of the Variables used in the Logit Model
Notation Name Definition
Explanatory Variables
Mother Education Education level of the mother
(dummy variables)
Moth_illiterateR Mother is illiterate = 1, if mother is illiterate
= 0, otherwise
Moth_elem Mother has elementary level of
education
= 1, if mother is having 1 to 8 years of
schooling
= 0, otherwise
Moth_sec Mother has secondary level of
education
= 1, if mother is having 9 to 10 years of
schooling
= 0, otherwise
Moth_higher Mother has senior secondary and
above level of education
= 1, if mother is having 11 or more years of
schooling
= 0, otherwise
Father Education Education level of the father
(dummy variables)
Fath_illiterateR Father is illiterate = 1, if father is illiterate
= 0, otherwise
Fath_elem Father has elementary level of
education
= 1, if father is having 1 to 8 years of schooling
= 0, otherwise
Fath_sec Father has secondary level of
education
= 1, if father is having 9 to 10 years of
schooling
= 0, otherwise
Fath_higher Father has senior secondary and
above level of education
= 1, if father is having 11 or more years of
schooling
= 0, otherwise
Moth_medexp Mother exposure to media
(dummy variable)
= 1, if mother listens radio or watches TV or
read news paper
= 0, if mother does not listen radio or watches
TV or read news paper
Moth_empw Socio‐economic empowerment
of the mother (dummy variable)
= 1, if mother alone take decision on different
socio‐economic aspects of the family
= 0, if mother alone does not take decision on
different socio‐economic aspects of the family
Household Economic
Status
Economic Status of the
Household (dummy variables)
PoorR Household is poor = 1, if the household is poor
= 0, otherwise
Middle_income Household is lying in the middle
income category
= 1, if the household is lying in the middle
income category
= 0, otherwise
Rich Household is rich = 1, if the household is rich
31
Notation Name Definition
= 0, otherwise
Moth_working Working status of the mother = 1, if the mother is working
= 0, otherwise i.e. if the mother is not working
Dwater_pipe Drinking water facility in the
household (dummy variable)
= 1, if the household is getting pipe drinking
water
= 0, otherwise i.e. if the household is using
tube well or drug well or surface water for
drinking purpose
Toilet_Facility Whether the household having
toilet or not
= 1, if the household is having toilet
= 0, otherwise i.e. the household is not having
toilet
Delivery_place Mother’s place of the delivery = 1, if institutional delivery
= 0, otherwise i.e. home delivery
Mother Age Age of the mother (dummy
variables)
Moth_less_20R Mother is less than 20 years old = 1, if the mother is less than 20 years old
= 0, otherwise
Moth_20_29 Mother is 20 to 29 years old = 1, if the mother is 20 to 29 years old
= 0, otherwise
Moth_30_39 Mother is 30 to 39 years old = 1, if the mother is 30 to 39 years old
= 0, otherwise
Moth_40_49 Mother is 40 to 49 years old = 1, if the mother is 40 to 49 years old
= 0, otherwise
Birth Order Birth order of the child (dummy
variables)
Birth_Order_1R First child of the mother = 1, if the birth order of the child is 1
= 0, otherwise
Birth_Order_2 Second child of the mother = 1, if the birth order of the child is 2
= 0, otherwise
Birth_Order_3 Third child of the mother = 1, if the birth order of the child is 3
= 0, otherwise
Birth_Order_4 Fourth child of the mother = 1, if the birth order of the child is 4
= 0, otherwise
Birth_Order5 Fifth child or more = 1, if the birth order of the child is 5 or more
= 0, otherwise
HH_size Household size Total members of the household
Caste Caste of the students (dummy
variables)
= 1, if the student belongs to scheduled castes
and scheduled tribes
= 0, otherwise (includes students belongs to
other backward classes and general category)
Dependent Variable
IMR Infant Mortality Rate (dummy
variable)
= 1, if the infant has died
= 0, otherwise i.e. if the infant is alive
Notes: R used as reference category in the logit estimations, results given in Table 4.
32
Table A2
Summary Statistics
Variables NOB Mean SD Min Max
Moth_illiterate 44599 0.45 0.50 0 1
Moth_elem 44599 0.21 0.41 0 1
Moth_sec 44599 0.21 0.40 0 1
Moth_higher 44599 0.13 0.34 0 1
Fath_illiterate 44121 0.26 0.44 0 1
Fath_elem 44121 0.22 0.42 0 1
Fath_sec 44121 0.30 0.46 0 1
Fath_higher 44121 0.21 0.41 0 1
Moth_medexp 44550 0.73 0.44 0 1
Moth_empw 43087 0.45 0.50 0 1
Poor 44601 0.38 0.49 0 1
Middle_income 44601 0.19 0.39 0 1
Rich 44601 0.43 0.49 0 1
Moth_working 44601 0.33 0.62 0 1
Caste 42593 0.70 0.46 0 1
Dwater_pipe 41972 0.42 0.49 0 1
Toilet_Facility 41922 0.48 0.50 0 1
Delivery_place 44601 0.45 0.49 0 1
Moth_less_20 44601 0.03 0.16 0 1
Moth_20_29 44601 0.65 0.48 0 1
Moth_30_39 44601 0.29 0.46 0 1
Moth_40_49 44601 0.03 0.17 0 1
Birth_Order_1 44601 0.32 0.47 0 1
Birth_Order_2 44601 0.28 0.45 0 1
Birth_Order_3 44601 0.16 0.37 0 1
Birth_Order_4 44601 0.10 0.30 0 1
Birth_Order5 44601 0.14 0.35 0 1
HH_size 44601 6.81 3.29 1 35
Note: The number of observations (NOB) is 44,601 except for some variables with missing information. Means and SD
are reported, which were corrected for the differences in sampling probabilities.
List of ISID Working Papers
156 The “Special Category State” Conundrum in Odisha, Nilmadhab Mohanty, October 2013
155 WP02: Medical Devices Manufacturing Industry in India: Market Structure, Import Intensity and Regulatory Mechanisms, ISID‐PHFI Collaborative Research Programme: Working Paper Series, Pritam Datta, Indranil Mukhopadhyay & Sakthivel Selvaraj
154 WP01: Changing Pattern of Public Expenditure on Health in India: Issues and Challenges, ISID‐PHFI Collaborative Research Programme: Working Paper Series, Shailender Kumar Hooda
153 WP2013/04: Currency Concerns under Uncertainty: Case of China, Sunanda Sen
152 WP2013/03: Structural Changes in Indian Private Corporate Sector, M.R. Murthy & K.V.K. Ranganathan
151 WP2013/02: Growth and Structural Changes in Indian Industry: Oraganised Sector, T.P. Bhat
150 WP2013/01: Economic Growth and Employment Linkages: The Indian Experience, T.S. Papola
149 WP2012/07: Employment Growth in the Post‐Reform Period, T.S. Papola
148 WP2012/06: Estimation of Private Investment in Manufacturing Sector and Determinants in Indian States, Jagannath Mallick
147 WP2012/05: Regional Disparities in Growth and Human Development in India, Satyaki Roy
146 WP2012/04: Social Exclusion and Discrimination in the Labour Market, T.S. Papola
145 WP2012/03: Changing Factor Incomes in Industries and Occupations: Review of Long Term Trends, Satyaki Roy
144 WP2012/02: Structural Changes in the Indian Economy: Emerging Patterns and Implications, T.S. Papola
143 WP2012/01: Managing Global Financial Flows at the Cost of National Autonomy: China and India, Sunanda Sen
142 WP2011/05: High Non‐Wage Employment in India: Revisiting the ‘Paradox’ in Capitalist Development, Satyaki Roy
141 WP2011/04: Trends and Patterns in Consumption Expenditure: A Review of Class and Rural‐Urban Disparities, Satyaki Roy
* Most of the working papers are downloadable from the institute’s website: http://isidev.nic.in/ or
http://isid.org.in/
Institute for Studies in Industrial Development4, Institutional Area Phase II, Vasant Kunj, New Delhi - 110 070
Phone: +91 11 2676 4600 / 2689 1111; Fax: +91 11 2612 2448E-mail: [email protected]; Website: http://isid.org.in ISID ISID
The Institute for Studies in Industrial Development (ISID), successor to the Corporate Studies Group (CSG), is a national-level policy research organization in the public domain and is affiliated to the Indian Council of Social Science Research (ICSSR). Developing on the initial strength of studying India’s industrial regulations, ISID has gained varied expertise in the analysis of the issues thrown up by the changing policy environment. The Institute’s research and academic activities are organized under the following broad thematic areas: Industrialization; Corporate Sector; Trade, Investment and Technology; Regulatory Mechanism; Employment; Public Health; Media Studies; and Other Issues.ISID has developed databases on various aspects of the Indian economy, particularly concerning industry and the corporate sector. It has created On-line Indexes of 203 Indian Social Science Journals (OLI) and 18 daily English Newspapers. More than one million scanned images of Press Clippings on diverse social science subjects are available online to scholars and researchers. These databases have been widely acclaimed as valuable sources of information for researchers studying India’s socio-economic development.
The Public Health Foundation of India (PHFI) is a public private initiative that has collaboratively evolved through consultations with multiple constituencies including Indian and international academia, state and central governments, multi & bi-lateral agencies and civil society groups. PHFI is a response to redress the limited institutional capacity in India for strengthening training, research and policy development in the area of Public Health.
Structured as an independent foundation, PHFI adopts a broad, integrative approach to public health, tailoring its endeavours to Indian conditions and bearing relevance to countries facing similar challenges and concerns. The PHFI focuses on broad dimensions of public health that encompass promotive, preventive and therapeutic services, many of which are frequently lost sight of in policy planning as well as in popular understanding.
About the PHFI
About the ISID
ISID-PHFI Collaborative Research Programme
PARENTAL EDUCATION AND INFANT MORTALITY IN INDIA:
Understanding the Regional Differences
November 2013
Working PaperNo: 157
Pradeep Kumar Choudhury
ISID-PHFI Collaborative Research Programme